Refining Question for Math Experts

oldbaldyone · October 2017

I'm looking to be able to work out the actual chances of refinement per try. Can define/someone explain the equation for this? I may be using the wrong terms for this explanation, give me some rope.

IE If I try to upgrade an enchant with a 3% success rate, what are the odds/chances of success each subsequent turn.

I fully understand that every time you try, you have 3% chance of success and that never changes. But, what are the "odds" I guess that after 50 tries for example, I still haven't had a success?

I know its not very useful for anything, but I'd love to know so i can have a better idea of the odds. If it could be an excel formula to work it out,even better

Thanks!

c3rb3r3 · October 2017

oldbaldyone said:
I fully understand that every time you try, you have 3% chance of success and that never changes. But, what are the "odds" I guess that after 50 tries for example, I still haven't had a success?

It's pure RNG, you can think you have better odds so that your mind remain sain, but the reality is after every fails it's like you restart from 0.

edit : to consider what my comrade said below, interesting.

plasticbat · October 2017

Computer RNG does not like flipping a coin in real life. It is just a sequence of number with a seed number feed into it. It is pseudo random and not random. If you have full control of the RNG as it only generates random number for you and nobody else, you should really get close to 3% chance in 100 tries or 3% in (say) 10000 tries.

If you serve the same seed number to RNG, the sequence of "random" number always has the same order. If this sequence of number is evenly distributed, the chance of the next number hit is higher if the last number failed.

I would say "every time you try, you have 3% chance of success and that never changes" is not true for pseudo random number. Roll a dice, yes. Computer RNG, no.

Of course, what I said is just a very simple RNG scenario, there are many factor to consider. For example, the algorithm may decide to apply a new seed very so often to mess it up like casino re-shuffle the deck of black jack when it reaches 3/4.

In addition, I don't think you will have full control of the RNG in this game (as the RNG will generate number for others) and I also suspect this game cache the RNG result.

thefabricant · October 2017

https://en.wikipedia.org/wiki/Binomial_distribution

If you are trying to work out the chance of having succeeded after 50 tries, it is 1-P(Fail).

P(Fail) = 0.97^50
1-P(Fail) = 0.78193462465

Which means that 78.19% of the time, you would have succeeded with 50 trials and 22% of the time, you would have failed and yeah, this explanation sucks, I am not in explaining mood atm.

namelesshero347 · October 2017

I think the posted success rate is FOS. I believe there are periods of time where "wins" are suppressed (if say the system thinks there has been too many winners recently), and where the RNG is juiced to produce more winners (if there were too few winners recently).

Its the only explanation to why I see so many results like needing 40+ wards for a 10% chance. Sadly, I rarely see the other end. Once I made two 3% upgrades with 5 wards.

The RNG is very streaky. Which is bad for players since getting caught in a losing streak can mean heavy losses to the button mashers. And most of the time when someone "wins", they stop to do something else, no chance to get a win streak.

plasticbat · October 2017

namelesshero347 said:

The RNG is very streaky. Which is bad for players since getting caught in a losing streak can mean heavy losses to the button mashers. And most of the time when someone "wins", they stop to do something else, no chance to get a win streak.

Actually, there are win streaks but usually are small things. In 2xRP (no more now), it is the time I open leadership bags. I had 6 to 7 green hit in sequence many times.
For XP rewards, I had a streak of 4 30K AD in a row. Need to do that quick.

oldbaldyone · October 2017

> @thefabricant said:
> https://en.wikipedia.org/wiki/Binomial_distribution
>
> If you are trying to work out the chance of having succeeded after 50 tries, it is 1-P(Fail).
>
> P(Fail) = 0.97^50
> 1-P(Fail) = 0.78193462465
>
> Which means that 78.19% of the time, you would have succeeded with 50 trials and 22% of the time, you would have failed and yeah, this explanation sucks, I am not in explaining mood atm.

This is what I was getting at and just needed to know the term. Thanks Sharp.

hammbo1969 · October 2017

I upgraded 3xR12 bondings today. Had 112 green wards and hit the first two in 20. Waiting to buy more to complete the third. there's no pattern to it - just pure HAMSTER RNG.

cilginordek · October 2017

(Fail Chance)^(Attempt Amount) gives you the odds of not succeeding.

minotaur2857 · October 2017

Streaky RNG is bad for players because if you fail you often try again immediately, if you succeed you usually do something else first.

Having just run out of pres wards because it took 71 attempts to get a 20% chance to work I don't believe a word of the alleged percentages. Also not one of the 40 or so enchants I upgraded at 5-30% chances where they started empty of RP so the first attempt filled the bar from scratch as well as trying to upgrade worked.

That said my research over tens of thousands of trials over several years suggests the numbers of attempts required are consistently 20-30% worse than they should be.

plasticbat · October 2017

minotaur2857 said:
Streaky RNG is bad for players because if you fail you often try again immediately, if you succeed you usually do something else first.

I guess my habit of "if it fails, I wait for many seconds; if I success, I hurry up; if I fail in a streak, I stop and do something else" helped me.

Refining Question for Math Experts

Comments